AI Research choices based on project cost?

Distant Worlds is a vast, pausable real-time, 4X space strategy game which models a "living galaxy" with incredible options for replayability and customizability. Experience the full depth and detail of large turn-based strategy games, but with the simplicity and ease of real-time, and on the scale of a massively-multiplayer online game. Now greatly enhanced with the new Universe release, which includes all four previous releases as well as the new Universe expansion!

Moderators: Icemania, elliotg

Post Reply
Bebop Cola
Posts: 250
Joined: Wed May 09, 2012 12:52 pm

AI Research choices based on project cost?

Post by Bebop Cola »

Does the AI choose between research projects based on relative cost?

For example, if I modify the race docs to give everyone a massive cost penalty for alternate planet type colonization research, will the AI shift its research priorities to cheaper projects instead at early game?

Also, do we know what the priority is for trading tech, specifically colonization tech? If the Gizureans meet the Sluken and hit off a friendship, what's the likelihood that they will immediately swap their native colonization techs as a way to avoid their individual research cost penalties considering that the colonization techs themselves wouldn't necessarily have a value increase?

Lastly, I've noticed that when using the editor to set tech levels back to starting tech, it appears that empires don't usually lose alternate colonization tech levels. I assume this is related to them having colonies on planet types different from their native type, effectively telling the game that their "native planet" at the point of the tech reset is any planet type they happen to have colonies for. I'll start a new game this evening and run a couple tests to confirm unless someone can do so before then.
Beag
Posts: 90
Joined: Wed May 23, 2012 1:22 pm

RE: AI Research choices based on project cost?

Post by Beag »

Good question. I´ve never ever received a tech offer, so maybe even the AI doesn´t do such deals between them? I did get offers to swap disputed bases, but never tech.
Bebop Cola
Posts: 250
Joined: Wed May 09, 2012 12:52 pm

RE: AI Research choices based on project cost?

Post by Bebop Cola »

I've had tech offers before. I usually try to keep a good relationship with the Kiadians, and they've offered me their Shadowghost before. I've also seen other empires get colonization tech earlier than I would expect them to. I'm assuming that they're trading for that tech with their friends/allies.

Ultimately, I want to delay or restrict the colonization rush to non-native planet types. Given my options for doing so, my concerns are as follows:

1. Making colonization techs 5 times more expensive(the limit, it seems, within the race files) will not adjust the AI's research priorities. Rather than bypassing colonization tech until later in the game in favor of making do with their native planet type and improving their other tech areas, they'll just try to muscle through an expensive colonization tech grind, stalling their other tech advancement.
2. If not #1, then they may adjust their trade preferences to prioritize trading their native tech(assuming they can) for that of their neighbors, bypassing my primary goal.

I can tolerate them conquering a non-native planet type and building colony ships from that planet that can access non-native planet types. At least, insofar as the majority population on that planet still consists of aliens. For example, if the Gizureans conquer a Quameno ocean world, they can initially build ocean colonizers that create Quameno colonies under the Gizurean banner. However, if Gizureans migrate to the Quameno world and end up the majority, they can no longer build ocean colonizers from that world. Instead, they produce their usual volcanic colonizers since the Gizureans are the majority.
Beag
Posts: 90
Joined: Wed May 23, 2012 1:22 pm

RE: AI Research choices based on project cost?

Post by Beag »

Another thing that could be fixed is how the AI simply doesn´t focus on weapons classes, instead researching a bit of everything. It´s yet another way the AI shoots at its own foot.

Was seriously considering editing race files so that they focus only in 2 or even one weapon class... How do I do that? And does it work or it will still research other stuff? Also, offering unique race tech is a bit... dumb. Zenox and Kiadian tech is potent.
Fishman
Posts: 795
Joined: Thu Apr 01, 2010 6:56 pm

RE: AI Research choices based on project cost?

Post by Fishman »

Offering unique race tech, in and of itself, is not necessarily "dumb". First, if you don't sell it, they're going to steal it sooner or later. Second, only a few races actually have really valuable tech: Kiadians, Ikkuro, Ugnari, and Haakonish. Even the basic versions of these items outclass the "public" version. Mortalen, Sluken, and Zenox technology is valuable also, but only when maxed out, which means it is very unlikely for it to ever mature in AI hands in time to make any real difference in your designs, as they will remain outclassed by the public domain versions unless maxed out, which the AI is in no hurry to do.

The other racial techs are mostly of limited-to-no value. The Dhayut drive prior to the new techtree USED to be The Best, AND you had it from the start of the game, but now it's outclassed, which means the Dhayut basically lost their only actual technical selling point, leaving them strictly for Ur-Quan fans only, as their engine technology is now an inferior dead-end. Quameno reactor was and remains crap. Shaktur firestorm is okay, but many options exist, and the Boskara aren't going to mature it anytime soon, so you're stealing it off the Shakturi. Ackdarian engine, was and remains worthless: Power efficiency is irrelevant when the engine performance is so bad. The fuel you save won't help when you're dead because your ship didn't move fast enough. Wekkarus PWC, same story as Shaktur Firestorm: Won't mature, immature versions outclassed, not an obvious must-have even when maxed.

So, basically, the top-4 racial techs are valuable, to the point where if they DON'T trade them, you will steal them, and those not in the top 4 are worthless and good only as bargaining chips. If that, for some of them.
Beag
Posts: 90
Joined: Wed May 23, 2012 1:22 pm

RE: AI Research choices based on project cost?

Post by Beag »

But isn´t stealing tech done by levels? I mean, first you must complete stealing the first version of the Zenox shield, then the second, then the third? Also, not all races will have good intelligence agents and many times, decent chances of sucess only happen in missions of 1 year, thus delaying tech steal a lot. Sometimes agents take ages to be created. Overall, I´d say tech stealing is balanced and not THAT easy. One thing is using humans, other is using Atuuk or other race that very seldom gets agents.

And all race specific weapons are potent, the problem again is that they research all weapons instead of focusing only on one or two. If Wekkarus rushed their wave weapons for example, they would be nasty. Instead, they research... ion weapons (derp derp).

Finally, the fact that in history mode the player gets so many advanced ships ahead of the AI is a serious balancing problem, so much that I don´t consider history mode balanced as it is now (a pity, because it´s the most fun).

EDIT: remembered another one. What must be edited to make colony ships only being able to be built with larger populations? It´s too easy nowadays to get independent or even AI planets and colonize stuff you previously couldn´t. That gives yet another advantage to the player. I want to be able to produce colony ships only if population higher than 500k. Which file must be edited?
Fishman
Posts: 795
Joined: Thu Apr 01, 2010 6:56 pm

RE: AI Research choices based on project cost?

Post by Fishman »

ORIGINAL: Beag

And all race specific weapons are potent, the problem again is that they research all weapons instead of focusing only on one or two. If Wekkarus rushed their wave weapons for example, they would be nasty. Instead, they research... ion weapons (derp derp).
I have not found this to actually be true. Also, you need to research Ion Weapons, otherwise you get raped by Ion Weapons because the Ion Defenses are in the Ion Weapons tree. In my games, the Wekkarus routinely are fielding level 2 waveblasters, but do not actually have that many "wrong" weapons. As you mentioned Ion Cannons, which, of course, is necessary because you need Ion Defenses, or you cease to have a functioning ship in short order, and the other tech often seen is Missiles, because you need missiles to have PDCs, or fighters eat you alive.
ORIGINAL: Beag

Finally, the fact that in history mode the player gets so many advanced ships ahead of the AI is a serious balancing problem, so much that I don´t consider history mode balanced as it is now (a pity, because it´s the most fun).
I haven't actually found this to be true. I have routinely seen AIs flying around with Historic ships. However, they tend not to actually realize what they're using, and therefore, get it lost in space. Plus, maybe they're doing the same thing I do with them: Scrapping them for research purposes, because they're so horribly made.
ORIGINAL: Beag

EDIT: remembered another one. What must be edited to make colony ships only being able to be built with larger populations?
Far as I know, you can't. But isn't this already in the game? There's definitely a minimum population needed to be able to deploy a colony ship. Additionally, good LUCK deploying a colony ship off a shithole backwater planet. The construction takes literally years.
ORIGINAL: Beag

It´s too easy nowadays to get independent or even AI planets and colonize stuff you previously couldn´t. That gives yet another advantage to the player. I want to be able to produce colony ships only if population higher than 500k. Which file must be edited?
Again, I don't think there is a file you can edit, but realistically, this is already sort of true: There definitely already *IS* a minimum population, and even a 500K world takes FOREVER to build that damn colony ship.

And no, I don't psychotically overdesign my colony ships, I actually strip them even more than the default designs, since they crash when landed and cannot be reused...so they clearly don't need stuff like "shields" or "armor", unless it's very advanced armor, in which case I give them a single plate for deflecting annoying pirate blasters, as pirate blasters can't penetrate UDA.
Beag
Posts: 90
Joined: Wed May 23, 2012 1:22 pm

RE: AI Research choices based on project cost?

Post by Beag »

Sure there is a minimum population for building colony ships, but my issue is with going after independents ASAP and using them to colonize stuff that would take ages normally (for example, getting a colony of space Barneys/dinos and colonize Ice planets that otherwise you would take much longer to be able, since Ice colonization is very expensive). The AI isn´t that smart to recognize that strategic potential. So what I intended to do was to make colony ships buildable only on planets with very high population, to brake such cheesy exploits.

Also, let me explain again my point about weapons. Specialization trumps diversification. Badly. Considering the standard start (one planet, level 0 tech), if I rush a tier 4 weapon (for example phaser lance) I will spank the AI that has tier 2 ion cannon, tier 2 Maxos, tier 2 interceptors (specially because fighters require a critical mass to do real damage and once I blow the base, the fighters are gone too). So the AI would be more challenging if it focus on no more than 3 weapons, preferably 2 (one close and one long range), else it could be exploited. But not the complete diversification it does now.

And the real advantage the player has isn´t the abandoned ships in systems, but damaged ships AND the depots that are in nebulae. Or that lovely Devastator.
Bebop Cola
Posts: 250
Joined: Wed May 09, 2012 12:52 pm

RE: AI Research choices based on project cost?

Post by Bebop Cola »

I think there would be some benefit to dedicating the colonization technology trees for each race. For example, any colonization technologies researched by the Keterov to colonize desert worlds would not be particularly applicable to the Naxillian efforts to do the same. While there might realistically be some crossover that could be gleaned, I don't expect that there is any game mechanism for accommodating the idea that stealing or buying Keterov desert world colonization could only ever give a certain percentage towards the Naxillian desert colonization project. As such, it might be best to just totally restrict colonization tech by race.

That said, doing so starts to tread on migration grounds. If the Naxillians can't make use of Keterov colonization techniques, how could they migrate to a Keterov world? Not sure how to answer that unless there is a way to restrict automatic migration to planets for which your empire does not have colonization tech for. Of course, then we have to wonder about cross-empire migration...
Beag
Posts: 90
Joined: Wed May 23, 2012 1:22 pm

RE: AI Research choices based on project cost?

Post by Beag »

Another option would be simply to not be able to build colony ships as a foreign race unless it is of the same race family.

Honestly, IMO there is too much tree-hugging in this game. Why would a nation actually pay for ships for an alien race to go out and expand? If I were a douchebag Boskara or Dhayut I would explore as much as I could foreign races, instead of allowing them to expand and chew resources the "master race" would use later. It´s related to the issue of conquered and foreign peoples being waaaay too passive. Since no one really hates each other, a space fly and space Barney can be friends and maybe even try to create cute Barney fly hybrids...
Bebop Cola
Posts: 250
Joined: Wed May 09, 2012 12:52 pm

RE: AI Research choices based on project cost?

Post by Bebop Cola »

ORIGINAL: Beag
Another option would be simply to not be able to build colony ships as a foreign race unless it is of the same race family.
You still run into issues such as with Humans/Kiadians being continental and Securans being desert, or Gizurean/Boskara being volcanic, Sluken being marsh, and Dhayut being desert.

Just because they're the same race family doesn't necessarily mean we can breath the same air, so to speak. Neither does it mean we would necessarily want to. Maybe the Securans are stinky and the Kiadians like their air mix high enough on oxygen to make humans loopy. Maybe the Sluken require a moisture level that rots the Dhayut exoskeleton.
Fishman
Posts: 795
Joined: Thu Apr 01, 2010 6:56 pm

RE: AI Research choices based on project cost?

Post by Fishman »

ORIGINAL: Bebop Cola
Of course, then we have to wonder about cross-empire migration...
I like the way that massive numbers of aliens will migrate to my planets...to be my slaves. This one baffles me, that they're migrating, literally by the billions, to my planets where I enslave them. Life must be really bad in their homelands if even being my slaves is a better life than where they come from.
ORIGINAL: Beag

Honestly, IMO there is too much tree-hugging in this game. Why would a nation actually pay for ships for an alien race to go out and expand?
Because you can't live there anyway, so sending those guys, who are either your citizens and therefore have the same rights as any of your other people, or your slaves, in which case they're there to prepare the planet for the arrival of their masters, makes perfect sense?
ORIGINAL: Beag

If I were a douchebag Boskara or Dhayut I would explore as much as I could foreign races, instead of allowing them to expand and chew resources the "master race" would use later.
I don't know if you've noticed this, but the Dhayut specifically LIKE doing that on purpose. Have you taken a look at their victory conditions lately? Yes, you are actually required to expand the quantity of foreignians you have beneath your...whatever, I guess. Do Dhayut have boots or thumbs? What exactly is the Dhayut symbol of oppression and tyranny, anyway? However, keep in mind that there is a sizeable differene between "chew" and "consume". I'm pretty sure the Dhayut don't mind having their slaves chew their food for them.
ORIGINAL: Beag

It´s related to the issue of conquered and foreign peoples being waaaay too passive.
Populations in general are pretty passive, whether they're conquered or yours. They're just a number. They don't actually have any apparent real interests or wants, just a number and happiness value.
Beag
Posts: 90
Joined: Wed May 23, 2012 1:22 pm

RE: AI Research choices based on project cost?

Post by Beag »

I´d rather see fixes to races´ passivity so that the choice would be expanding fast and facing rebellions or even independence attempts sometimes or expanding slow and not recruting troops from foreign races; but failing that, I´d accept limitations of colony ships only to races of the same family. It´s up to the developers to create such options, through (something like "advanced population management", in case the average player thinks that adds too much complexity to the standard game).
Post Reply

Return to “Distant Worlds 1 Series”